AI improvement prompts

Hvussu tað riggar

Tvey dátustykki koma saman við klikki:

• Findingin — heiti, lýsing, prógv, bøting, CWE — er longu lødd við frágreiðingini.
• Frameworkið hjá tínari codebase — funnið úr discovery.tech-fingerprint findings hjá skanningini (Next.js, React, Vue, Django, Express, Rails, Laravel, Flask). Fellur aftur til eina generiska uppskrift, tá einki framework verður funnið (agent-prompturin biður tá LLM finna tað úr repo-støðuni).

Templates live in lib/scanner/fix-prompts.ts. The registry has check-specific guidance for crawlability, search presentation, semantic content, structured data, media, performance, accessibility, forms, mobile/i18n, runtime, owner journeys, and repo-quality issues. For everything else, the existing remediation field on the finding becomes the generic recipe.

Hvussu prompturin sær út

Fix the "Hero image is lazy-loaded and missing dimensions" finding on /pricing.

Issue: The largest above-the-fold image is marked loading="lazy" and has
no explicit width/height. That can delay LCP and create layout shift.

Codebase context: Next.js.

Recommended fix:
Use next/image or the existing image component with explicit width/height,
responsive sizes, meaningful alt text, and priority/fetchPriority for the
first major visual on the page.

Constraints:
- Don't break existing tests; run the test suite after the change.
- Match the codebase's existing style and lint config.
- Add a brief comment explaining the performance reasoning only where the
  fix would otherwise look arbitrary.
- If the fix needs a new dependency, install it via the project's
  package manager (npm / pnpm / pip / bundle / composer).

Reference: Core Web Vitals / Largest Contentful Paint guidance.

Stuðlað frameworks

Vit vísa framework-serstakar snippets fyri:

• Next.js, React, Vue, Nuxt, Svelte (framendi)
• Express, Fastify (Node.js bakendi)
• Django, Flask (Python)
• Ruby on Rails
• Laravel (PHP)
• ASP.NET Core (ætlað, fellur aftur til generiskt í dag)

Framework-finning er best-effort. Vit sniffa __NEXT_DATA__-tags, __NUXT__, hash cookies (laravel_session), X-Powered-By-headerar og nøkur onnur signal. Um tú koyrir eitt sergjørt framework, fellur prompturin aftur til generisku uppskriftina, og agenturin finnur tað úr tínum package.json.

Brúka tað frá tínum AI-agenti

Um tú hevur bundið MCP tænastaran inn, er sami promptur útsettur sum slash command. Frá Claude Desktop:

/fixweb-fix finding_id=550e8400-e29b-41d4-a716-446655440000

The renderer looks up the finding, detects the framework from the parent scan when available, renders the templated prompt, and injects it into your conversation as the user message. No round-trip to our Claude API; templates are pure and free.

Hví vit ikki kalla Claude fyri hvørt klikk

Við launch umhugsaðu vit at kalla Anthropic API fyri hvørt klikk fyri at fínpussa promptin við codebase-konteksti. Tað gjørdu vit ikki, tí:

• Agenturin, sum brúkarin setur inn í, hevur longu codebase-kontekst — tey brúka Cursor / Claude Desktop við opnum repo.
• Templating fyri hvørt (check × framework) dekkar ~80% av virðinum uttan nakran kostnað fyri hvørt klikk.
• Ein “Refine with AI for my codebase”-opt-in kundi seinni latið API'ið fire'a, um brúkarar vilja tað. Í dag: nei.